Synthesis, structure, and biological evaluation of novel N- and O-linked diinositols

Article abstract of DOI:10.1021/ja0205378

Several O-and N-linked inositols and/or aminoinositols have been prepared by iterative opening of epoxides and aziridines derived from homochiral cyclohexadiene cis-diols. The three inositols and their intermediate conduritols (conduramines) were tested a

Full text of DOI:10.1021/ja0205378

Published on Web 08/10/2002
Synthesis, Structure, and Biological Evaluation of Novel N-
and O-Linked Diinositols
Bernhard J. Paul, Jerremey Willis, Theodore A. Martinot, Ion Ghiviriga,†
Khalil A. Abboud,‡ and Tomas Hudlicky*
Contribution from the UniVersity of Florida, Department of Chemistry, PO Box 117200,
GainesVille, Florida 32611
Received April 15, 2002
Abstract: Several O-and N-linked inositols and/or aminoinositols have been prepared by iterative opening
of epoxides and aziridines derived from homochiral cyclohexadiene cis-diols. The three inositols and their
intermediate conduritols (conduramines) were tested against several glycosidases (R- and â-glucosidase,
R- and â-galactosidase, R- and â-mannosidase) in an assay that measured the rate of hydrolysis of
p-nitrophenolglycosides rather than the concentration of p-nitrophenolate. Somewhat surprisingly, the best
inhibition was seen against â-galactosidase with several of the compounds. The inositols linked through
oxygen or nitrogen were subjected to calcium binding studies performed in NMR experiments. Detailed
analysis of the title compounds by NMR spectroscopy has been performed, and full assignments were
made. One of the attendant benefits of the preparation of these compounds has been expressed in the
design and synthesis of new salen catalysts whose effectiveness has been compared with Jacobsen’s
catalyst in the epoxidation of styrene.
atoms or groups.8 Along with polyhydroxylated alkaloids,9 most
such derivatives show modest to excellent inhibition of the
Introduction
Investigations of the biological function of sugars have
witnessed enormous progress over the last several years, and
cellular processes such as recognition events are now more
understood, allowing for the development of carbohydrate-based
drugs.1 Unnatural derivatives of carbohydrates have been at the
forefront of bioorganic chemistry research primarily because
of their ability to mimic the structure of monosaccharides and
hence serve as inhibitors of glycosidases. Since these enzymes
are involved in a variety of different biological processes, such
research naturally leads to potential applications in the treatment
of diabetes,2 viral infections,3 and cancer.3
common glycosidases that would be involved in the initial steps
of cell penetration by the viral material. Similarly, disaccharides
and even oligosaccharides have been modified10 and studied
further for their role in the inhibition of glycosidases.
Several years ago we disclosed the preparation of a series of
simple oligomers containing the configuration of L-chiro-inositol
as shown by the general structure 1, Figure 1.11 The initial reason
for investigating such compounds was based on the assumption
that O-linked oligomers of this type may play a role in insulin
mediation12 and at the same time not be sensitive to the gastric
environment, thus being potential candidates for orally admin-
istered medicines.
Recent literature presents many examples of simple carbo-
hydrates in which the oxygen has been replaced with carbon
(pseudosugars),4 nitrogen (iminosugars),5,6 sulfur,7 and other
A number of oligomers have been prepared,13 and some of
these compounds proved to have interesting metal chelating
properties14 and displayed tertiary structures in solid state. In
an effort to investigate compounds that would mimic the
* Address correspondence to this author. E-mail: hudlicky@chem.ufl.edu.
† To whom correspondence regarding NMR spectroscopy should be
addressed.
‡ To whom correspondence regarding X-ray crystallography should be
addressed.
(8) Lillelund, V. H.; Jensen, H. H.; Liang, X.; Bols, M. Chem. ReV. 2002,
102, 515-553.
(1) (a) Dwek, R. A. Chem. ReV. 1996, 96, 683-720. (b) Yarema, K. J.; Bertozzi,
C. R. Curr. Opin. Chem. Biol. 1998, 2, 49-61. (c) Sears, P.; Wong, C.-H.
Angew. Chem., Int. Ed. 1999, 38, 2300-2324. (d) Alper, J. Science 2001,
291, 2338-2343. (e) Gruner, S. A. W.; Locardi, E.; Lohof, E.; Kessler, H.
Chem. ReV. 2002, 102, 491-514.
(9) (a) Asano, N.; Nash, R. J.; Molyneux, R. J.; Fleet, G. W. J. Tetrahedron:
Asymmetry 2000, 11, 1645-1680. (b) Watson, A. A.; Fleet, G. W. J.; Asano,
N.; Molyneux, R. J.; Nash, R. J. Phytochemistry 2001, 56, 265-295.
(10) Liu, L.; McKee, M.; Postema, M. H. D. Curr. Org. Chem. 2001, 5, 1133-
1167.
(2) Truscheit, E.; Hillebrand, I.; Junge, B.; Mu¨ller, L.; Puls, W.; Schmid, D.
Prog. Clin. Biochem. Med. 1988, 7, 17-99.
(11) Paul, B. J.; Martinot, T. A.; Willis, J.; Hudlicky, T. Synthesis 2001, 952-
956.
(3) Zitzmann, N.; Mehta, A. S.; Carrouee, S.; Butters, T. D.; Platt, F. M.;
McCauley, J.; Blumberg, B. S.; Dwek, R. A.; Block, T. M. Proc. Natl.
Acad. Sci. 1999, 96, 11878-11882.
(12) (a) Ley, S. V.; Yeung, L. L. Synlett 1992, 997-998. (b) Reddy, K. K.;
Falck, J. R.; Capdevila, J. Tetrahedron Lett. 1993, 34, 7869-7872.
(13) (a) Hudlicky, T.; Abboud, K. A.; Entwistle, D. A.; Fan, R.; Maurya, R.;
Thorpe, A. J.; Bolonick, J.; Myers, B. Synthesis 1996, 897-911. (b)
Desjardins, M.; Lallemand, M.-C.; Hudlicky, T.; Abboud, K. A. Synlett
1997, 728-730. (c) Desjardins, M.; Lallemand, M.-C.; Freeman, S.;
Hudlicky, T.; Abboud, K. A. J. Chem. Soc., Perkin Trans. 1 1999, 621-
628.
(4) Ferrier, R. J.; Middleton, S. Chem. ReV. 1993, 93, 2779-2831.
(5) Stuetz, A. E. Iminosugars as Glycosidase Inhibitors; Wiley-VCH: Wein-
heim, 1998.
(6) Berecibar, A.; Grandjean, C.; Siriwardena, A. Chem. ReV. 1999, 99, 779-
844.
(7) Robina, I.; Vogel, P.; Witczak, Z. J. Curr. Org. Chem. 2001, 5, 1177-
1214.
(14) Hudlicky, T.; Abboud, K. A.; Bolonick, J.; Maurya, R.; Stanton, M. L.;
Thorpe, A. J. J. Chem. Soc., Chem. Commun. 1996, 1717-1718.
9
10416
J. AM. CHEM. SOC. 2002, 124, 10416-10426
10.1021/ja0205378 CCC: $22.00 © 2002 American Chemical Society

Novel N- and O-Linked Diinositols
A R T I C L E S
Acylation of 8 with acetic anhydride under rather vigorous
conditions afforded acetamide 10 (88%). Hydrolysis of the ester
groups in 10 under basic conditions (NaOMe, MeOH) followed
by acidic cleavage of the acetonides (TFA/THF/H2O) provided
the acetamide-linked diconduritol 11 in 85% yield for two steps.
To obtain the fully hydroxylated species, dimer 10 was
subjected to OsO4-catalyzed dihydroxylation (OsO4 cat./NMNO/
acetone/H2O) followed by protection of the resulting diols as
their acetonides (DMP, H+) to give dimer 12 in 60% yield for
two steps. Hydrolysis of the esters (NaOMe, MeOH) followed
by removal of the acetonides under acidic conditions (HCl,
MeOH) furnished the acetamide-linked diinositol 13 in 57%
yield (two steps).
For the synthesis of amine-linked diinositol 16, compound 8
was treated with acetic anhydride at low temperature to achieve
acylation of the alcohols only, leaving the sterically hindered
secondary amine untouched. Acylation of the latter with
trifluoroacetic anhydride provided trifluoroacetamide 14 (78%
for two steps). Oxidation with a catalytic amount of OsO4
followed by protection of the resulting diols as their acetonides
gave the fully hydroxylated species 15 (53%). Cleavage of the
esters as well as the trifluoroacetamide under basic conditions
(NaOMe, MeOH) followed by acidic hydrolysis (HCl, MeOH)
provided amine-linked diinositol 16 as its hydrochloride in 75%
yield. For the synthesis of diaminoinositol dimer 17, diol 5 was
converted to vinyl aziridine 18 according to a known proce-
dure.17,18 Ring opening of 18 was achieved with ammonia under
Yb(OTf)3 catalysis followed by treatment of the intermediate
amine with another equivalent of vinyl epoxide 6 to afford amine
19 in 78-94% yield. Conversion of the alcohol into its acetate
provided amine 20 (93%), which was subsequently converted
to trifluoroacetamide 21 (92-98%). Removal of the halogens
under radical conditions (AIBN, Bu3SnH) afforded alkene 22
in 56-73% yield. Osmylation of the double bonds followed
by protection of the resulting diols as their acetonides afforded
the fully oxygenated species 23 in 26-61% yield. Hydrolysis
of the acetate and trifluoroacetamide under basic conditions
(57%), removal of the sulfonamide with Na in liquid ammonia
(84%), and acidic hydrolysis of the acetonides afforded diinositol
17 as its hydrochloride salt in 77% yield. For the preparation
of oxygen-linked diinositol 24, we envisioned a convergent
strategy. Vinyl epoxide 6 was opened under basic conditions
to provide diol 25 in 63% yield. Reaction of 25 with vinyl
aziridine 18 under BF3‚Et2O catalysis afforded a mixture of two
ethers 26a (52%) and 26b (18%), along with several byproducts
(vide infra). Removal of the halogens under radical conditions
gave 27a (92%) and 27b (86%). After hydroxylation of the
double bonds and protection of the resulting diols as acetonides,
both isomers converged into a single compound 28 (65%).
Deprotection of the tosylamide with Na in liquid ammonia
followed by hydrolysis of the acetals under acidic conditions
provided free dimer 24 as its hydrochloride salt in 52% yield.
Figure 1. Structural comparison of N- and O-linked (amino)inositol
oligomers to natural occurring antibiotics.
connectivity found in naturally occurring antibiotics such as
validoxylamine 2 and fortimicin 3, Figure 1,15 we wished to
prepare and examine O- and N-linked oligomers that would
contain such a motif. To this end, N- and O-linked diinositols
and their precursory conduritols (or conduramines) would be
targeted in order that their biological properties may be
examined against common glycosidases. In this article we report
the details of the syntheses of compounds of the general type
4, the preparation of some of which has been disclosed in a
preliminary form and which lend themselves to further oligo-
merizations through the interaction of Y with either epoxides
or aziridines.11 In addition, the results of testing of these
compounds against six glycosidases are provided. The future
potential of these compounds as three-dimensional templates
for transfer of C-O chiral connectivity to C-C asymmetry is
also indicated.
Synthesis
Bromocyclohexadiene-cis-diol (5)16 was obtained by
whole-cell fermentation of bromobenzene with E. coli
JM109(pDTG601), an organism that expresses toluene dioxy-
genase (TDO) and was further converted to vinyl epoxide 6.17
Reaction of 6 with ammonia in the presence of Yb(OTf)3
resulted in rapid ring opening of the epoxide to give an amino
alcohol as the intermediate product, which was without isolation
treated with another equivalent of vinyl epoxide 6 to afford
dimer 7 in a one-pot process in 78% yield, Scheme 1. Removal
of the bromine atoms under radical conditions (AIBN, Bu3SnH)
afforded alkene 8 in 83% yield. Treatment of 8 with HCl in
MeOH yielded amine-linked diconduritol 9 as its hydrochloride
(84%).
Upon examining the byproducts of the Lewis-acid-mediated
coupling, we were also able to identify compounds 26c, 29,
and 30 in the reaction mixture (Figure 2).
(15) (a) Ogawa, S.; Ogawa, T.; Chida, N.; Toyokuni, T.; Suami, T. Chem. Lett.
1982, 749-752. (b) Ogawa, S.; Ogawa, T.; Iwasawa, Y.; Toyokuni, T.;
Chida, N.; Suami, T. J. Org. Chem. 1984, 49, 2594-2599.
(16) (a) Gibson, D. T.; Hensley, M.; Yoshioka, H.; Mabry, T. J. Biochemistry
1970, 9, 1626-1630. (b) Hudlicky, T.; Luna, H.; Barbieri, G. L.; Kwart,
L. D. J. Am. Chem. Soc. 1988, 110, 4735-4741. (c) Hudlicky, T.; Stabile,
M. R.; Gibson, D. T.; Whited, G. M. Org. Synth. 1999, 76, 77-85.
(17) Hudlicky, T.; Tian, X. R.; Ko¨nigsberger, K.; Rouden, J. J. Org. Chem.
1994, 59, 4037-4039.
The presence of the cis-configured dimer 26c was somewhat
surprising and indicated that this material was formed by an
(18) (a) Tian, X. R.; Hudlicky, T.; Ko¨nigsberger, K. J. Am. Chem. Soc. 1995,
117, 3643-3644. (b) Hudlicky, T.; Tian, X. R.; Ko¨nigsberger, K.; Maurya,
R.; Rouden, J.; Fan, B. J. Am. Chem. Soc. 1996, 118, 10752-10765.
9
J. AM. CHEM. SOC. VOL. 124, NO. 35, 2002 10417

A R T I C L E S
Paul et al.
Scheme 1 a
a
Reagents and conditions: (i) DMP, H+; (ii) mCPBA, CH2Cl2; (iii) NH3, Yb(OTf)3; (iv) 6, dioxane, ∆; (v) AIBN, Bu3SnH, THF; (vi) HCl, MeOH; (vii)
Ac2O, pyr. DMAP, dioxane, ∆; (viii) NaOMe, MeOH; (ix) THF, TFA, H2O (2:2:1); (x) OsO4, NMNO, acetone, H2O; (xi) Ac2O, pyr. DMAP, CH2Cl2, 0 °C;
(xii) TFAA, pyr. DMAP, dioxane ∆.
SN1-type mechanism. Surprisingly, the relative ratio of these
products (26a:26b:26c:29:30 ) 2:1:1:1:1) proved to be generally
unaffected by changes in temperature, catalyst concentration,
and solvent, except for a higher percentage of 30 in the reaction
mixture at high catalyst concentrations.
pairs in the DQCOSY spectra recorded in DMSO-d6. For dimers
16 and 16a, having identical inositol units, complete
1H and
13C assignments were made on the basis of the 1H-1H and 1H-
13C correlations seen in the GHMQC, GHMBC, DQCOSY, and
NOESY spectra. The size of the coupling constants between
the CH protons in positions 1 through 6 confirmed the
stereochemistry of the inositol rings. The chemical shifts in
DMSO-d6 at 25 °C were referenced to TMS and are given in
Table 1. Variable-temperature spectra in the interval 25-75 °C
afforded the temperature coefficients of the 1H chemical shifts,
also reported in Table 1. The relative exchange rates of the OH
protons with water, measured in the NOESY spectrum, paral-
leled the temperature coefficients, indicating that the latter are
indicative of the accessibility of the OH groups to the solvent
molecules. The NOEs are not very useful in the study of the
conformation in solution of these molecules because of their
Structure Assignment by NMR
The structure assignments of the N- and O-linked oligomers
were inferred initially from first principles of reactivity of
epoxides and aziridines (assuming trans-diaxial opening of the
strained three-membered rings with nucleophiles) and confirmed
by detailed analysis of 1H NMR spectra and, in some cases, by
X-ray crystallography.
Severe overlap of the ring protons in the dimers having
different inositol moieties prevented the assignment of the proton
signals; however they displayed the expected 10 H(C)-H(X)
9
10418 J. AM. CHEM. SOC. VOL. 124, NO. 35, 2002

Novel N- and O-Linked Diinositols
A R T I C L E S
Scheme 2 a
a
Reagents and conditions: (i) DMP, H+; (ii) PhINTs, Cu(acac)2, CH3CN; (iii) NH3, Yb(OTf)3, (iv) 6, dioxane, ∆; (v) Ac2O, pyr. DMAP, CH2Cl2; (vi)
TFAA, pyr. DMAP, ∆; (vii) AIBN, Bu3SnH, THF, (viii) OsO4, NMNO, acetone, H2O; (ix) NaOMe, MeOH; (x) Na, NH3 (l); (xi) HCl, MeOH.
Scheme 3 a
a
Reagents and conditions: (i) KOH, DMSO, ∆; (ii) BF3.Et2O, 18; (iii) AIBN, Bu3SnH, THF; (iv) OsO4, NMNO, acetone, H2O; (v) DMP, H+; (vi) Na,
NH3 (l); (vii) HCl, MeOH.
symmetry. The NOESY spectrum of 16, however, revealed fast
exchange between the NH2 protons at 8.10 ppm and the OH
The NMR data for compounds 26a-c raises intriguing
questions about the outcome of the ring opening of aziridine
18 by diol 25. Both the 1H-1H vicinal couplings and the 1H-
13C long-range couplings confirm that 26b results from the
reaction of the hydroxyl at C-3 of diol 25, while 26a and 26c
are the products of the reaction of the C-4 hydroxyl in 25. The
values of the coupling constants given in Figure 1 demonstrate
that 26a and 26b are indeed the expected products of trans
aziridine ring opening, but in 26c strongly support a structure
+
protons in position 4, at 5.55 ppm. Corroborating this with a
small temperature coefficient for the OH in position 2 (5.24
ppm), indicative of intramolecular hydrogen bonding, one can
infer for 16 a solution conformation in which rotation about
the C-N bonds brings the OH’s in position 2 closer, within a
distance appropriate for mutual hydrogen bonding, and the OH’s
in position 6 closer to the ammonium protons.
9
J. AM. CHEM. SOC. VOL. 124, NO. 35, 2002 10419

A R T I C L E S
Paul et al.
Table 2. 1H and 13C Chemical Shifts and Temperature
Coefficients of the 1H Chemical Shifts for Compounds 26a-c
26a
26b
26c
δH
dδH/dT
δC
δH
dδH/dT
δC
δH
dδH/dT
δC
position (ppm) (ppb/K) (ppm) (ppm) (ppb/K) (ppm) (ppm) (ppb/K) (ppm)
1
120
120.4
122.7
2
3
4
5
6
7
8
9
1′
2′
3′
4′
5′
6′
7′
8′
9′
6.36
4.15
3.67
4.07
4.61 -0.5
1.43 -0.1
1.33 -0.2
0
0
0.5
0.5
133.6 6.37
77.7 3.96
57.2 3.62
77.1 4.15
77.2 4.61
27.8 1.41
26.2 1.34
111.4
0.4 133.2 6.47 -0.3 132.2
2
0
77.5 3.92
56.7 3.53
9.7 76.2
2.3 55
0.4 76
4.3
-0.7 74.8
0.1 77.3 4.63 -1.2 77.4
0.7 27.9 1.33 -2.7 27.7
0.6 26.4 1.36 -1.2 26.2
118.8
118.4
110.5
117.8
119.2
6.22 -0.1 134.3 5.98
3.4 132.7 6.26 -1.7 135.9
3.97
3.44
4.17
4.63
1.43
1.54
0.5
0.8
1.9
0.1
0
69.3 3.8
0.9 78.3 4.12 -1.7 70.5
81.4 3.48
2.7 72.7 3.18
6
83.8
1.3 77.4
77.4 4.62 -0.7 77.2 4.61 -0.2 77.6
25.9 1.57 0.4 28.4 1.39 2.7 26
77.4 4.04 -0.1 77.3 4.03
Figure 2. Byproducts of the BF3‚Et2O-catalyzed coupling reaction.
0
28
111.2
1.42 -0.1 26
1.39 -0.2 27.6
111.4
111
OH 2.76 -5.2
NH 5.34 -6.5
1′′
2.6
4.76 -11
-4.1
4.01 -19.3
6.58 -19.8
137.8
139.1
138
2′′
3′′
4′′
5′′
7.83
0.2 127.7 7.78
0.6 127.4 7.84 -0.8 127.7
7.29 -0.2 129.4 7.3
-
2.42
0
129.9 7.32 -0.8 129.9
144 144.1
22.2 2.42 -0.3 21.4
-
143.3
0
21.7 2.44
0
Figure 3. Position numbering for compounds 16 and 16a.
originating in aziridine of 26b to the coupling constants on the
rings originating in diol of 26a-c demonstrates that 26b is the
product of trans ring opening (with H3, H4, and H5 all
pseudoaxial) and that the coupling constants do not depend
significantly on the substitution pattern of the ring. The values
Table 1. 1H and 13C Chemical Shifts and Temperature
Coefficients of the 1H Chemical Shifts for Compounds 16 and 16a
16
16a
δHCH
δC
δHOH dδHOH/d T δHCH
δC
δH
dδHOH/d T
(ppb/K)
3
position (ppm) (ppm) (ppm)
(ppb/K)
(ppm) (ppm) (ppm)
of the vicinal coupling constants of H3 in 26c, JH2-H3 ) 5.5
Hz and 3JH3-H4 ) 3.7 Hz, clearly demonstrate that this hydrogen
is equatorial and that 26c is the product of the cis ring opening.
The NOEs seen in the NOESY spectrum of 26c recorded in
CDCl3 at -25 °C support this conclusion: H3′ and H5′ display
the NOE expected for 1,3-diaxial protons, while H3 and H5 do
not. Similarly, 26a,b display the NOEs between H3 and H5
and between H3′ and H5′. The NOESY spectrum of 26a
indicates that the solution conformation is very similar to that
observed in the solid state (Figure 5).
NOEs between H2 and H8′, H3 and H8′, and H2′ and H2′′
demonstrate that of all the possible conformations arising from
rotation about the C3-O-C4′ bond, 26a adopts one in which
C9′ is close to C2. This is also confirmed by the NOEs of
comparable size between H2 and H3 and between H2 and H4′
and by a larger NOE of H4′ with H2 than with H4. A larger
NOE of the OH with H2 than with H2′ indicates that the OH is
oriented toward the tosyl group, but a small value for the
temperature coefficient of the OH chemical shift suggests lack
of hydrogen bonding. A coupling constant of 7.6 Hz for the
NH, together with larger NOEs to H3 and H5 than to H4,
indicates that the NH is antiperiplanar to H4. The phenyl group
prefers a position on the same side of the molecule as C7, as
suggested by NOEs between H2′′ and H7.
1
2
3
4
5
6
3.72 72.5 5.2
3.95 68.1 5.24
3.53 63.7
3.75 70.1 5.55
3.49 71.5 4.77
3.68 72.2 4.96
-4.8
-1.6
3.72 72.4 4.49
3.65 71.9 5.68
3.43 65.7
-6.7
-4.2
-4.4
-6.4
-5.2
3.47 73.8 5.62
-6
-7.6
-5.4
3.5
71.6 4.16
3.67 72.7 4.46
in which the ring opening occurred cis, indicative of an SN1-
like substitution of a cation or its ion pair.
The chemical shifts given in Table 1 have been determined
in CDCl3 at 25 °C. The chemical shift assignment was based
on 1H-1H couplings seen in selective decoupling experiments,
and on one-bond and long-range 1H-13C couplings seen in
GHMQC and GHMBC spectra, respectively.
The assignment of methyls of the acetonide groups was made
based on the NOEs seen in the NOESY spectra, namely, strong
NOEs between H8 and H5, H6 and between H7′ and H5′, H6′.
The temperature coefficients of the chemical shifts have been
measured in CDCl3, in the temperature interval -25 to 55 °C.
The consistency of the coupling constants JH4-H5, JH4′-H5′,
3JH5-H6, and 3JH5′-H6′ in compounds 26a-c clearly demonstrates
that both cyclohexene rings in all three compounds are in the
same conformation, one in which the hydrogens in positions 4,
5, 4′, 5′ are pseudoaxial. Vicinal coupling constants of 8-9 Hz
for the protons in positions 3′, 4′, and 5′ in compounds 26a-c
confirm that these protons are all pseudoaxial. A value of ca. 2
3
3
The NOESY spectrum of 26b, in CDCl3 at -15 °C, displays
an NOE between H2 and H4′, which indicates that of all the
conformations arising from rotation about the C3-O-C4′ bond,
26b adopts the one in which the OH and the NH are anti. NOEs
of similar magnitude of H2 and H2′ with H3 and H3′ suggest
3
Hz for JH2′-H3′ also indicates that H3′ is axial, a position in
which the dihedral angle H2′-C2′-C3′-H3′ is close to 90°.
The similarity of the vicinal coupling constants on the ring
9
10420 J. AM. CHEM. SOC. VOL. 124, NO. 35, 2002

Novel N- and O-Linked Diinositols
A R T I C L E S
Figure 6. Comparison of the structures of 24 and 33.
indicate that in 26c both of the exchangeable protons are
involved in hydrogen bonding. The solvent, CDCl3, is not a
hydrogen bond acceptor; therefore the hydrogen bonds have to
be intramolecular, and of all the conformations arising from
rotation about the C3-O-C4′ bond, 26c has to adopt at low
temperature the one in which the OH and the NH are syn. The
NOESY spectrum of 26c at -25 °C displays the NOEs expected
for such a conformation, i.e., H2-H7′, H2-H5′, OH-NH, H5-
OH, NH-H3′, and OH-H2′′.
The NH is antiperiplanar to H4 (3JNH-H4 ) 8.0 Hz) and forms
a hydrogen bond to the oxygen in the hydroxyl group, across
the eight-member ring H‚‚‚O-C3′-C4′-O-C3-C4-N, a
picture confirmed by five large NOEs of the NH with OH, H4′,
H3, H4, and H5. Large NOEs of H7′ with H2 and H3 agree
with this hydrogen bonding, but only for the structure in which
the substituents in positions 3 and 4 are cis. The OH proton
displays a coupling of 4.2 Hz with H3′ and NOEs with H2′,
H3′, and H5; its most likely location is intercalated between
C3′-H3′ and C3′-C2. The coupling constants of the OH and
NH do not vary significantly with temperature, indicating that
the formation of the hydrogen bond produces only a small
change in the orientation of these protons relative to their vicinal
partners. The NOEs of H2′′ with H3, H4, and H4′ and the NOE
of H3′′ with H7′ all demonstrate that the tosyl moiety is on the
same side of the eight-member ring H‚‚‚O-C3′-C4′-O-C3-
C4-N as H4′. The protons having a large positive temperature
coefficient of the chemical shifts, H3, H4, H4′, and H7′, are in
the same region of space, which probably is the shielding region
of the tosyl ring in the most stable conformation. Large
temperature coefficients for the chemical shifts of H7 (negative)
and H7′ (positive) are more consistent with reorientation of the
tolyl group upon the formation of a hydrogen bond between
the OH hydrogen and an oxygen in the sulfonic moiety than
with rotation about the C3-O-C4′ bond upon the formation
of the NH‚‚‚O bond.
Figure 4. Position numbering and vicinal coupling constants (in Hz) for
compounds 26a-c.
Figure 5. Crystal structure of dimer 26a as determined by X-ray diffraction.
that the bridge C3-O-C3′ is twisted as to bring H3′ closer to
H2. Proton H2′′ displays NOEs with H2′, H4′, and H7′, which
confirm that the tosyl group is anti to the OH with respect to
the C3-O-C3′ junction and on the same side of the cyclo-
hexene ring C1′-C6′ as H3′. The OH displays a vicinal coupling
constant of 2.0 Hz and the only NOE with H4′; therefore it is
most likely in one of the two gauche positions to this one. The
NH has a vicinal coupling of 8.0 Hz and NOEs with H3, H4,
and H5, consistent with the NH being antiperiplanar to H4. The
temperature coefficients of the chemical shifts are smaller in
26b than in 26c. The protons experiencing larger values are
H3, H2′, and H4′; most likely the variation of the chemical shifts
with the temperature is due to changes in the angle H3-C3-
O-C3′-H3′. The hydroxyl proton has a small temperature
coefficient, and most likely is not involved in any hydrogen
bonding.
Calcium Binding Affinity Study
Our earlier report on the Ca2+ sequestering properties
associated with amino diinositol 3314 prompted us to investigate
the Ca2+ binding potential of all three oligomers, but especially
24, as it is configurationally identical to 32 except for one extra
hydroxyl group. The binding affinity toward Ca2+ was studied
for compounds 16, 17, 24, and 33 by monitoring the change in
chemical shifts of the CH protons in deuterated water, upon
titration with a D2O solution of known ratio of CaCl2 and acetic
acid. The acetic acid was used to calculate the Ca2+/inositol
ratio from the proton spectra. A plot of chemical shifts versus
concentration showed a small variation of the chemical shifts
(less than 0.03 ppm for a ratio Ca2+/inositol of 5) for compounds
16, 24, and 33. The change in the chemical shifts during titration
of CH’s 1-6 in these compounds also paralleled the temperature
The exchangeable protons, NH and OH, display significant
differences in chemical shifts and, most importantly, in the
temperature coefficients of the chemical shifts between com-
pounds 26c and 26b. Large negative temperature coefficients
9
J. AM. CHEM. SOC. VOL. 124, NO. 35, 2002 10421

A R T I C L E S
Paul et al.
Figure 7. Calcium interaction studies of dimer 17. Top spectrum is without Ca2+, bottom one contains excess calcium ions.
Figure 8. Inhibition of R-mannosidase by DIM (31).
coefficients of the corresponding OH’s; therefore they merely
reflect the change in the ionic strength of the solution and no
Ca2+ binding occurs. In the case of compound 17, a net change
in chemical shift (0.1 to 0.2 ppm) is observed for what would
be the protons nearest the nitrogen in the molecule, namely,
protons in positions 3 and 4, which were identified by their
two large couplings, signifying an axial proton with two axial
neighbors. On the other hand, the protons assigned on the
“periphery” of the molecule barely shift at all. This is clearly
indicative of the molecule binding with calcium.
enzymes was very labor intensive, especially since we encoun-
tered problems with the reproducibility of our data using this
protocol. To ensure a faster and less error-prone inhibition
evaluation, we set our focus to modifying the current protocol
to allow for a more rapid screening. (For a more detailed
description see the Experimental Section.) All compounds were
screened against the corresponding enzymes at or close to their
optimum pH value, and the liberated p-nitrophenolate was
monitored at 400 nm without a basic quench. We used a
multicell spectrometer, which allowed the monitoring of the
(19) For examples see: (a) Saul, R.; Molyneux, R. J.; Elbein, A. D. Arch.
Biochem. Biophys. 1984, 230, 668-675. (b) Fleet, G. W. J.; Smith, P. W.;
Evans, S. V.; Fellows, L. E. J. Chem Soc., Chem. Commun. 1984, 1240-
1241. (c) Dale, M. P.; Ensley, H. E.; Kern, K.; Sastry, K. A. R.; Byers, L.
D. Biochemistry 1985, 24, 3530-3539. (d) Daniels, L. B.; Coyle, P. J.;
Chiao, Y.-B.; Glew, R. H.; Labow, R. S. J. Biol. Chem. 1981, 256, 3004-
3013.
Glycosidase Inhibition Studies
For the evaluation of the glycosidase inhibition properties of
inositol dimers, we first utilized the conventional assay.19
However, the screening of all compounds toward several
9
10422 J. AM. CHEM. SOC. VOL. 124, NO. 35, 2002

Novel N- and O-Linked Diinositols
A R T I C L E S
Figure 9. Glycosidase inhibition of inositol oligomers.
Figure 10. Inhibition of â-galactosidase.
enzymatic reaction at different inhibitor concentrations over a
period of time (usually 20 min). For active compounds we
observed a decrease of the initial rate with increasing inhibitor
concentration. When plotting the reaction rates (normalized to
the initial rate ) 100%) against the inhibitor concentrations,
inhibition curves were generated from which the corresponding
IC50 values could be obtained either graphically or by math-
ematical means.
Results and Discussion
To verify the reproducibility and accuracy of the assay, we
compared the results of the new testing procedure with literature
9
J. AM. CHEM. SOC. VOL. 124, NO. 35, 2002 10423

A R T I C L E S
Paul et al.
Scheme 4 a
a
Reagents and conditions: (i) NaOMe, MS 3 Å, MeOH, EtOH, ∆; (ii) Mn(OAc)2‚4H2O, EtOH, PhCH3, ∆; (iii) air, then NaCl (sat.); (iv) DMP, H+.
Table 3. Glycosidase Inhibition of Inositol Oligomers
inhibitor tested (IC , µM)
50
enzyme
13
11
16
9
24
17
â-galactosidase
â-glucosidase
R-mannosidase
R-glucosidase
R-galactosidase
â-mannosidase
980
n.I.a
n.I.
370
n.I.
n.I.
100
n.I.
n.I.
1350
n.I.
n.I.
750
n.I.
n.I.
n.I.
n.I.
n.I.
620
n.I.
n.I.
1400
n.I.
n.I.
1120
n.I.
n.I.
1750
n.I.
n.I.
n.I.
n.I.
n.I.
n.I.
n.I.
n.I.
Figure 11. Comparison of the configuration of 13, 11, and â-D-galactose
(32).
that will be confirmed or rejected when all monomeric units of
13 and 11 as well as the corresponding trimeric compounds
are subjected to the same inhibition studies.
a n.I. ) less than 50% inhibition at an inhibitor concentration of 2000
µM.
Inositol-Based Salen Complexes. During the synthesis of
diamino derivative 19 we discovered a surprisingly efficient
procedure for the synthesis of trans-1,2-diaminotetrahydroxy-
cyclohexanes. A preliminary account of the TBAF-catalyzed
opening of a tosylaziridine with toluenesulfonamide has been
published,21 and an initial investigation of the role of this
compound in the design of (potentially) water-soluble catalysts
for Jacobsen epoxidation has been made.
values obtained by the traditional way. As a test substrate we
chose 1,4-dideoxy-1,4-imino-D-mannitol (DIM) (31), a powerful
inhibitor of R-mannosidase (Figure 8).
The literature reports an IC50 value for DIM against R-
mannosidase of 0.5 µM ([S] ) 0.33 mM; pH ) 4.5), whereas
our assay produced a value of 0.35 µM ([S] ) 2.5 mM; pH )
6.0).20
By using the protocol described in this article, we obtained
the inhibition data shown in Table 3 and Figure 9.
It was surprising to find that almost all oligomers were active
against â-galactosidase, despite the fact that the peripheral
substitution of these compounds does not correspond to the
configuration of the natural D-galactose.
When comparing the structure of L-chiro-aminoinositol 13
and conduramine-F with â-D-galactose, it is evident that they
do not possess the same hydroxylation pattern, except at three
of the hydroxylated carbons. We may speculate that hydroxyl
groups from both inositol rings as well as the acetamide oxygen
might participate in the binding to the enzyme, a speculation
Thus the salen catalysts 34 and 35 have been prepared as
shown in Scheme 4.
Previously reported21 3,4-diamino-3,4-dideoxy-L-chiro-inosi-
tol (36) was treated with aldehyde 37 and base to afford salen
complex 38.22 The latter could be converted to manganese
complex 34 by modifying Jacobsen’s procedure.23 Compound
38 was protected to afford salen-type species 39, which was
converted to protected catalyst 35 by following the same
protocol.
(21) Paul, B. J.; Hobbs, E.; Buccino, P.; Hudlicky, T. Tetrahedron Lett. 2001,
42, 6433-6435.
(22) Belekon, Y.; Ikonov, N.; Moscalenko, M.; North, M.; Orlova, S.; Tararov,
V.; Yashkina, L. Tetrahedron: Asymmetry 1996, 7, 851-855.
(23) Larrow, J. F.; Jacobsen, E. N.; Gao, Y.; Hong, Y.; Nie, X.; Zepp, C. M. J.
Org. Chem. 1994, 59, 1939-1942.
(20) Because of the low equilibrium concentration of p-nitrophenolate at pH )
4.5, we were unable to perform the assay at this low pH.
9
10424 J. AM. CHEM. SOC. VOL. 124, NO. 35, 2002

Novel N- and O-Linked Diinositols
A R T I C L E S
Figure 12. C-O to C-C connectivity translation.
Sodium phosphate buffer (25mM) pH ) 6.0 for R-mannosidase,
R-galactosidase
Both catalysts 34 and 35 were compared to Jacobsen’s
original catalyst in the chiral epoxidation of styrene.24,25
Jacobsen’s catalyst gave S-styrene oxide in an enantiomeric
excess (ee) of almost 60% in this reaction at -78 °C, whereas
catalysts 34 and 35 afforded the R-configured product with an
ee of 16% and 27%, respectively. We were also surprised to
find that the four hydroxyl groups in catalyst 34 did not enhance
its water solubility at all. Further endeavors in this field will
focus on attaching polar groups (e.g., sulfates) on the alcohol
moieties to increase the water solubility.
Sodium phosphate buffer (25mM) pH ) 5.5 for â-mannosidase
The inhibitor concentration was kept around 1 mg/ mL. All tests
were run at 37 °C, except for the R-mannosidase assay, which was
performed at 25 °C. The substrate concentration was 5 mM except for
the R- and â-mannosidase assay, where it was 2.5 mM. The enzyme
concentration was adjusted to produce a slope of approximately 0.02-
0.025 except for the â-mannosidase assay, where it was adjusted to
give a slope of 0.005-0.01. The final volume (adjusted with the
corresponding buffer) was 1000 µL in all cases.
To a solution of varying amounts of inhibitor (usually 0-400 µL)
in the corresponding buffer solution was added 100 µL of the enzyme
solution, except for one vial that was run as a blank to correct for the
autohydrolysis of the substrate. All samples were preincubated at the
corresponding temperature for 30 min, after which the reaction was
started by addition of the substrate (400 µL). The absorption was then
recorded over a period of 20 min by continually scanning all samples.
The slopes obtained for the individual samples were then plotted against
the inhibitor concentration to obtain the usual inhibition curves.
Aziridine-Alcohol Couplings: Typical Procedure. Borontrifluo-
ride diethyl ether complex (24.7 µL, 250 µmol) in dry CH2Cl2 (15 mL)
was added to a stirred solution of aziridine 18 (1.0 g; 2.50 mmol) and
diol 25 (0.66 g; 2.50 mmol) at -50 °C. The mixture was allowed to
warm to -10 °C and stirred at that temperature for 20 min. A saturated
solution of NaHCO3 (20 mL) was added, and mixture was extracted
with CH2Cl2 (3 × 30 mL). The combined organic layers were dried
(MgSO4) and concentrated under reduced pressure. The products of
the reaction (26a-c, 29, 30) were separated by flash chromatography
(3:1 hexanes/ethyl acetate).
Conclusion
The N- and O-linked oligomers have been synthesized and
subjected to glycosidase inhibition studies. For one of these
compounds the precursory conduramine oligomers has been
found more active, by an order of magnitude, than the
corresponding fully hydroxylated analogue. This observation
is in agreement with postulated models for the transition states
in the hydrolysis of the glycosidic linkage: the cyclohexene
motif of the conduritol (conduramine) resembles more closely
the tetrahydropyranyl cation. Somewhat surprising was the
finding that all compounds showed, albeit modest, activity
against â-galactosidase. The configuration of its corresponding
substrate is not reinforced in any of the tested compounds.
These preliminary results bode well for further investigation
aimed at determining which structural units are responsible for
increase or decrease of antiglycosidic activities.
The compounds themselves are also suited for a host of
material or molecular recognition studies. Apart from the
opportunities in catalyst design for compounds containing the
trans-diamino cyclohexane motif, these structures may be used
as templates for potential transfer of chirality from C-O bonds
to C-C bonds by a number of processes such as free radical
cascades, Heck-type cascades, or Grubb’s metathesis bond
formation of oxygen-tethered functional groups. One such
possibility is outlined in Figure 12.
5,6-Bis[(3,5-di-tert-butyl-2-hydroxybenzylidene)amino]cyclohexane-
1,2,3,4-tetraol (38). To a solution of hydrochloride 36 (150 mg; 0.597
mmol), aldehyde 37 (308 mg; 1.3147 mmol), and crushed, activated
molecular sieves (3 Å) in a 1:1 mixture of EtOH and MeOH (15 mL
total) was added a 0.2 M solution of NaOMe in MeOH (12 mL), upon
which the reaction mixture turned bright yellow. The solution was
heated to reflux for 15 h and then allowed to cool to room temperature
and concentrated in vacuo. The residue was redissolved in CH2Cl2 (20
mL) and filtered through a plug of silica and evaporated again. The
residue was loaded on a flash-chromatography column and eluted with
hexanes/ethyl acetate (1:1) to obtain 38 as a yellow foam (292 mg;
Further results in the investigation of chemistry, structure,
biological activity, and catalytic or template roles of these
oligomers will be reported in due course.
25
80%): [R]D +206.6 (c 0.60, CHCl); IR (film on NaCl) 3403, 2956,
2924, 2869, 1626, 913, 743 cm-1; 1H NMR (CDCl3, 300 MHz) δ 8.35
Experimental Section26
(26) All nonaqueous reactions were performed using standard techniques for
the exclusion of moisture and air. Methylene chloride and dioxane were
distilled from calcium hydride, whereas ether and THF were dried over
sodium/benzophenone. Thin-layer chromatography was performed on
Silicycle plates and flash chromatography using Natland 200-400 mesh
silica gel. Melting points were recorded on a Hoover Unimelt apparatus
and are uncorrected. IR spectra were recorded on a Perkin-Elmer FT-IR or
on a Perkin-Elmer Spectrum One FT-IR spectrometer. 1H and 13C NMR
spectra were recorded on a Varian Gemini (300 MHz), a Varian VXR (300
MHz), a Mercury 300 (300 MHz), and an Inova 500 (500 MHz) instrument.
All chemical shifts are referenced to TMS or residual undeuterated solvent.
Optical rotations were measured on a Perkin-Elmer 341 instrument. All
combustion analyses were performed by Atlantic Microlab, Norcross, GA.
Mass spectra were recorded by the analytical division at the University of
Florida, Gainesville.
General Procedure for the Inhibition Assays. The following
buffers were used for the tests:
Sodium phosphate buffer (25mM) pH ) 7.3 for â-galactosidase
Sodium phosphate buffer (25mM) pH ) 6.8 for â-glucosidase,
R-glucosidase
(24) Palucki, M.; Pospisil, P. J.; Zhang, W.; Jacobsen, E. N. J. Am. Chem. Soc.
1994, 116, 9333-9334.
(25) Because of the easy availability of both enantiomers of styrene oxide, we
chose styrene as a substrate, despite its moderate suitability as a substrate
for the Jacobsen epoxidation.
9
J. AM. CHEM. SOC. VOL. 124, NO. 35, 2002 10425

A R T I C L E S
Paul et al.
(s, 2H), 7.33 (d, J ) 2.4 Hz, 2H), 7.02 (d, J ) 2.4 Hz, 2H), 4.20-
117.9, 109.3, 77.7, 75.5, 71.2, 35.1, 34.2, 31.6, 29.6, 28.3, 25.8; HRMS
calcd for C42H63N2O6 (M + H)+ 691.4686, found 691.4683. Anal. Calcd
for C42H62N2O6‚0.5H2O: C, 72.07; H, 9.07; N, 4.00. Found: C, 71.91;
H, 8.98; N, 3.79.
4.4.32 (m, 4H), 3.64-3. 74 (m, 2H), 1.40 (s, 18H), 1.22 (s, 18H); 13
C
NMR (CDCl3, 75 MHz) δ 169.8, 158.1, 140.6, 136.8, 127.7, 126.8,
117.8, 71.9, 71.3, 70.2, 35.2, 34.3, 31.6, 29.6; HRMS calcd for
C36H55N2O6 (M + H)+ 611.4060, found 611.4062. Anal. Calcd for
C36H54N2O6: C, 70.79; H, 8.91. Found: C, 70.25; H, 8.95.
5,6-Bis[(2-hydroxy-3,5-dimethylbenzylidene)amino]cyclohexane-
1,2:3,4-di-O-isopropylidene-1,2,3,4-tetraol Manganese Complex (35).
In a two-neck flask equipped with a reflux condenser was dissolved
Mn(OAc)2‚4H2O (64 mg; 0.260 mmol) in EtOH (3 mL). The mixture
was heated to reflux, after which a solution of ligand 549 (60 mg;
0.868 mmol) in toluene (3 mL) was added dropwise over a 5 min period.
The mixture was allowed to reflux for 45 min, after which air was
bubbled through the solution for 5 h. After that time, no more ligand
39 could be detected by TLC monitoring. Heating and air were
discontinued, and the reaction mixture was quenched with brine (15
mL) and allowed to cool to room temperature. To the solution was
added toluene (15 mL), the layers were separated, and the organic layer
was washed with H2O (2 × 5 mL) and brine (1 × 5 mL). After the
solution was dried (Na2SO4) and concentrated, a brown residue was
obtained, which was dissolved in CH2Cl2 (1 mL). To this solution was
added heptane (10 mL), which caused precipitation after a few minutes.
The crystals were removed by filtration, the mother liquor was
evaporated, and the residue was subsequently dissolved again to obtain
three more fractions. The combined precipitates were dried at high
vacuum to obtain crude 35 as a brown solid (21 mg; 31%). No further
purification of 35 was performed: [R]D30 +141.5 (c 0.050, CHCl3); IR
(film on NaCl) 3435, 1644, 1616, 1532, 1260 cm-1; HRMS calcd for
C42H60MnN2O6 (M - Cl)+ 743.3832, found 743.3826. Anal. Calcd for
C42H60ClMnN2O6: C, 64.73; H, 7.76; N, 3.59. Found: C, 43.69; H,
5.52; N, 1.98.
5,6-Bis[(3,5-di-tert-butyl-2-hydroxybenzylidene)amino]cyclohexane-
1,2,3,4-tetraol Manganese Complex (34). In a three-neck flask
equipped with a reflux condenser was dissolved Mn(OAc)2‚4H2O (300
mg; 1.223 mmol) in EtOH (10 mL). The mixture was heated to reflux,
after which a solution of ligand 47 (249 mg; 0.407 mmol) in toluene
(10 mL) was added dropwise over a 5 min period. The mixture was
allowed to reflux for 1.5 h, after which air was bubbled through the
solution for 2 h. After that time, no more ligand 38 could be detected
by TLC monitoring. Heating and air were discontinued, and the reaction
mixture was quenched with brine (15 mL) and allowed to cool to room
temperature. To the solution was added toluene (50 mL), the layers
were separated, and the organic layer was washed with H2O (2 × 20
mL) and brine (1 × 20 mL). After the solution was dried (Na2SO4)
and concentrated, a brown residue was obtained, which was dissolved
in CH2Cl2 (1 mL). To this solution was added heptane (10 mL), which
caused precipitation after a few minutes. The crystals were removed
by filtration, the mother liquor was evaporated, and the residue was
subsequently dissolved again to obtain three more fractions. The
combined precipitates were dried at high vacuum to obtain 34 as a
26
brown solid (93 mg; 33%): [R]D +830 (c 0.013, CHCl3); IR (film
for C36H52MnN2O6 (M - Cl)+ 663.3206, found 663.3225. Anal. Calcd
for C36H52ClMnN2O6‚1.5H2O: C, 59.54; H, 7.63; N, 3.86. Found: C,
59.34; H, 7.33; N, 3.74.
5,6-Bis[(2-hydroxy-3,5-dimethylbenzylidene)amino]cyclohexane-
1,2:3,4-di-O-isopropylidene-1,2,3,4-tetraol (39). Ligand 38 (62 mg;
0.101 mmol) was dissolved in CH2Cl2 (5 mL) and DMP (200 µL). To
this solution was added a spatula tip of p-TsOH, and the mixture was
stirred at room temperature for 20 h. Additional DMP (200 µL) and
p-TsOH (1 spatula tip) was added, and the mixture was allowed to stir
an additional 10 h. After that, the reaction was quenched by adding
excess solid Na2CO3. The mixture was stirred for 5 min, filtered through
a plug of Celite, washed with CH2Cl2, and evaporated. The residue
was purified by flash chromatography (gradient elution: starting with
hexanes, ending with hexanes/ethyl acetate, 10:1) to obtain 39 as a
Acknowledgment. The authors are grateful to the following
agencies for generous financial support: The National Science
Foundation (CHE-9615112, CHE-9910412), the U.S. Environ-
mental Protection Agency (R- 826113), TDC Research Founda-
tion, Inc., TDC Research, Inc., the McKnight Foundation
(Fellowship for J.W.), and the University of Florida (fellowship
for TM). The authors would like to thank Prof. Jon Stewart for
helpful discussions and for the access to his spectrometer.
Skillful assistance of Pablo Buccino and Elizabeth Hobbs in
the preparation of starting materials is appreciated.
26
yellow oil (52 mg; 73%): [R]D +60.2 (c 1.0, CHCl3); IR (film on
NaCl) 2957, 2871, 1738, 1630, 1598, 1468, 1441, 1382, 1370, 1362,
1
1273, 1250, 1060 cm-1; H NMR (CDCl3, 300 MHz) δ 8.33 (s, 2H),
Supporting Information Available: Crystallographic data
(CIF). This material is available free of charge via the Internet
at http://pubs.acs.org.
7.32 (d, J ) 1.9 Hz, 2H), 7.03 (d, J ) 1.9 Hz, 2H), 4.61 (d, J ) 5.2
Hz, 2H), 4.48 (dd, J ) 7.1, 5.2 Hz, 2H), 3.44 (dd, J ) 5.2, 2.1 Hz,
2H), 1.58 (s, 6H), 1.40 (s, 6H), 1.39 (s, 18H), 1.25 (s, 18H); 13C NMR
(CDCl3, 75 MHz) δ 168.7, 158.2, 140.3, 136.6, 132.7, 127.5, 126.7,
JA0205378
9
10426 J. AM. CHEM. SOC. VOL. 124, NO. 35, 2002